Skip to main content

Table 1 Statistics on the number of clusters for various clustering options compared to Corset

From: Corset: enabling differential gene expression analysis for de novoassembled transcriptomes

  Chicken Human Yeast
  Trinity Oases Trinity Oases Trinity Oases
Contigs 335,377 540,933 107,389 239,426 7,353 27,013
Trinity Clusters (Max.) 230,924 (302)   73,258 (91)   6,690 (45)  
Oases Clusters (Max.)   87,639 (93,103)   55,746 (16,881)   3,140 (5,987)
CD-HIT-EST Clusters (Max.) 282,285 (81) 202,636 (116) 90,115 (29) 96,965 (74) 7,117 (8) 5,586 (39)
Corset Clusters (Max.) 91,653 (290) 67,826 (208) 43,663 (90) 38,476 (59) 3,796 (45) 4,324 (65)
  1. Shown are the number of contigs (bold), number of clusters and the maximum number of contigs in a cluster (in parentheses). Corset removes contigs that have less than 10 reads mapping to them by default, and hence has the least number of clusters in 5 out of 6 assemblies. This makes the final list of clusters more manageable, with no detriment to the final DGE results. Oases grossly over-clusters as shown by the maximum contigs in a cluster.